Listen to this post, minus the robot

LXB Voice has used Polly’s standard text-to-speech engine since 2022. It now uses the neural engine on every install where Voice is turned on.

The two engines work differently. Standard voices use concatenative synthesis, which stitches together phonemes of recorded speech. As AWS puts it, “the inevitable variations in speech and the techniques used to segment the waveforms limits the quality of speech.” The neural engine takes a sequence-to-sequence approach that AWS says yields …